Merge tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-nextgrafted

Pull networking updates from Jakub Kicinski:
 "Core:

   - Add dedicated kmem_cache for typical/small skb->head, avoid having
     to access struct page at kfree time, and improve memory use.

   - Introduce sysctl to set default RPS configuration for new netdevs.

   - Define Netlink protocol specification format which can be used to
     describe messages used by each family and auto-generate parsers.
     Add tools for generating kernel data structures and uAPI headers.

   - Expose all net/core sysctls inside netns.

   - Remove 4s sleep in netpoll if carrier is instantly detected on
     boot.

   - Add configurable limit of MDB entries per port, and port-vlan.

   - Continue populating drop reasons throughout the stack.

   - Retire a handful of legacy Qdiscs and classifiers.

  Protocols:

   - Support IPv4 big TCP (TSO frames larger than 64kB).

   - Add IP_LOCAL_PORT_RANGE socket option, to control local port range
     on socket by socket basis.

   - Track and report in procfs number of MPTCP sockets used.

   - Support mixing IPv4 and IPv6 flows in the in-kernel MPTCP path
     manager.

   - IPv6: don't check net.ipv6.route.max_size and rely on garbage
     collection to free memory (similarly to IPv4).

   - Support Penultimate Segment Pop (PSP) flavor in SRv6 (RFC8986).

   - ICMP: add per-rate limit counters.

   - Add support for user scanning requests in ieee802154.

   - Remove static WEP support.

   - Support minimal Wi-Fi 7 Extremely High Throughput (EHT) rate
     reporting.

   - WiFi 7 EHT channel puncturing support (client & AP).

  BPF:

   - Add a rbtree data structure following the "next-gen data structure"
     precedent set by recently added linked list, that is, by using
     kfunc + kptr instead of adding a new BPF map type.

   - Expose XDP hints via kfuncs with initial support for RX hash and
     timestamp metadata.

   - Add BPF_F_NO_TUNNEL_KEY extension to bpf_skb_set_tunnel_key to
     better support decap on GRE tunnel devices not operating in collect
     metadata.

   - Improve x86 JIT's codegen for PROBE_MEM runtime error checks.

   - Remove the need for trace_printk_lock for bpf_trace_printk and
     bpf_trace_vprintk helpers.

   - Extend libbpf's bpf_tracing.h support for tracing arguments of
     kprobes/uprobes and syscall as a special case.

   - Significantly reduce the search time for module symbols by
     livepatch and BPF.

   - Enable cpumasks to be used as kptrs, which is useful for tracing
     programs tracking which tasks end up running on which CPUs in
     different time intervals.

   - Add support for BPF trampoline on s390x and riscv64.

   - Add capability to export the XDP features supported by the NIC.

   - Add __bpf_kfunc tag for marking kernel functions as kfuncs.

   - Add cgroup.memory=nobpf kernel parameter option to disable BPF
     memory accounting for container environments.

  Netfilter:

   - Remove the CLUSTERIP target. It has been marked as obsolete for
     years, and we still have WARN splats wrt races of the out-of-band
     /proc interface installed by this target.

   - Add 'destroy' commands to nf_tables. They are identical to the
     existing 'delete' commands, but do not return an error if the
     referenced object (set, chain, rule...) did not exist.

  Driver API:

   - Improve cpumask_local_spread() locality to help NICs set the right
     IRQ affinity on AMD platforms.

   - Separate C22 and C45 MDIO bus transactions more clearly.

   - Introduce new DCB table to control DSCP rewrite on egress.

   - Support configuration of Physical Layer Collision Avoidance (PLCA)
     Reconciliation Sublayer (RS) (802.3cg-2019). Modern version of
     shared medium Ethernet.

   - Support for MAC Merge layer (IEEE 802.3-2018 clause 99). Allowing
     preemption of low priority frames by high priority frames.

   - Add support for controlling MACSec offload using netlink SET.

   - Rework devlink instance refcounts to allow registration and
     de-registration under the instance lock. Split the code into
     multiple files, drop some of the unnecessarily granular locks and
     factor out common parts of netlink operation handling.

   - Add TX frame aggregation parameters (for USB drivers).

   - Add a new attr TCA_EXT_WARN_MSG to report TC (offload) warning
     messages with notifications for debug.

   - Allow offloading of UDP NEW connections via act_ct.

   - Add support for per action HW stats in TC.

   - Support hardware miss to TC action (continue processing in SW from
     a specific point in the action chain).

   - Warn if old Wireless Extension user space interface is used with
     modern cfg80211/mac80211 drivers. Do not support Wireless
     Extensions for Wi-Fi 7 devices at all. Everyone should switch to
     using nl80211 interface instead.

   - Improve the CAN bit timing configuration. Use extack to return
     error messages directly to user space, update the SJW handling,
     including the definition of a new default value that will benefit
     CAN-FD controllers, by increasing their oscillator tolerance.

  New hardware / drivers:

   - Ethernet:
      - nVidia BlueField-3 support (control traffic driver)
      - Ethernet support for imx93 SoCs
      - Motorcomm yt8531 gigabit Ethernet PHY
      - onsemi NCN26000 10BASE-T1S PHY (with support for PLCA)
      - Microchip LAN8841 PHY (incl. cable diagnostics and PTP)
      - Amlogic gxl MDIO mux

   - WiFi:
      - RealTek RTL8188EU (rtl8xxxu)
      - Qualcomm Wi-Fi 7 devices (ath12k)

   - CAN:
      - Renesas R-Car V4H

  Drivers:

   - Bluetooth:
      - Set Per Platform Antenna Gain (PPAG) for Intel controllers.

   - Ethernet NICs:
      - Intel (1G, igc):
         - support TSN / Qbv / packet scheduling features of i226 model
      - Intel (100G, ice):
         - use GNSS subsystem instead of TTY
         - multi-buffer XDP support
         - extend support for GPIO pins to E823 devices
      - nVidia/Mellanox:
         - update the shared buffer configuration on PFC commands
         - implement PTP adjphase function for HW offset control
         - TC support for Geneve and GRE with VF tunnel offload
         - more efficient crypto key management method
         - multi-port eswitch support
      - Netronome/Corigine:
         - add DCB IEEE support
         - support IPsec offloading for NFP3800
      - Freescale/NXP (enetc):
         - support XDP_REDIRECT for XDP non-linear buffers
         - improve reconfig, avoid link flap and waiting for idle
         - support MAC Merge layer
      - Other NICs:
         - sfc/ef100: add basic devlink support for ef100
         - ionic: rx_push mode operation (writing descriptors via MMIO)
         - bnxt: use the auxiliary bus abstraction for RDMA
         - r8169: disable ASPM and reset bus in case of tx timeout
         - cpsw: support QSGMII mode for J721e CPSW9G
         - cpts: support pulse-per-second output
         - ngbe: add an mdio bus driver
         - usbnet: optimize usbnet_bh() by avoiding unnecessary queuing
         - r8152: handle devices with FW with NCM support
         - amd-xgbe: support 10Mbps, 2.5GbE speeds and rx-adaptation
         - virtio-net: support multi buffer XDP
         - virtio/vsock: replace virtio_vsock_pkt with sk_buff
         - tsnep: XDP support

   - Ethernet high-speed switches:
      - nVidia/Mellanox (mlxsw):
         - add support for latency TLV (in FW control messages)
      - Microchip (sparx5):
         - separate explicit and implicit traffic forwarding rules, make
           the implicit rules always active
         - add support for egress DSCP rewrite
         - IS0 VCAP support (Ingress Classification)
         - IS2 VCAP filters (protos, L3 addrs, L4 ports, flags, ToS
           etc.)
         - ES2 VCAP support (Egress Access Control)
         - support for Per-Stream Filtering and Policing (802.1Q,
           8.6.5.1)

   - Ethernet embedded switches:
      - Marvell (mv88e6xxx):
         - add MAB (port auth) offload support
         - enable PTP receive for mv88e6390
      - NXP (ocelot):
         - support MAC Merge layer
         - support for the the vsc7512 internal copper phys
      - Microchip:
         - lan9303: convert to PHYLINK
         - lan966x: support TC flower filter statistics
         - lan937x: PTP support for KSZ9563/KSZ8563 and LAN937x
         - lan937x: support Credit Based Shaper configuration
         - ksz9477: support Energy Efficient Ethernet
      - other:
         - qca8k: convert to regmap read/write API, use bulk operations
         - rswitch: Improve TX timestamp accuracy

   - Intel WiFi (iwlwifi):
      - EHT (Wi-Fi 7) rate reporting
      - STEP equalizer support: transfer some STEP (connection to radio
        on platforms with integrated wifi) related parameters from the
        BIOS to the firmware.

   - Qualcomm 802.11ax WiFi (ath11k):
      - IPQ5018 support
      - Fine Timing Measurement (FTM) responder role support
      - channel 177 support

   - MediaTek WiFi (mt76):
      - per-PHY LED support
      - mt7996: EHT (Wi-Fi 7) support
      - Wireless Ethernet Dispatch (WED) reset support
      - switch to using page pool allocator

   - RealTek WiFi (rtw89):
      - support new version of Bluetooth co-existance

   - Mobile:
      - rmnet: support TX aggregation"

* tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1872 commits)
  page_pool: add a comment explaining the fragment counter usage
  net: ethtool: fix __ethtool_dev_mm_supported() implementation
  ethtool: pse-pd: Fix double word in comments
  xsk: add linux/vmalloc.h to xsk.c
  sefltests: netdevsim: wait for devlink instance after netns removal
  selftest: fib_tests: Always cleanup before exit
  net/mlx5e: Align IPsec ASO result memory to be as required by hardware
  net/mlx5e: TC, Set CT miss to the specific ct action instance
  net/mlx5e: Rename CHAIN_TO_REG to MAPPED_OBJ_TO_REG
  net/mlx5: Refactor tc miss handling to a single function
  net/mlx5: Kconfig: Make tc offload depend on tc skb extension
  net/sched: flower: Support hardware miss to tc action
  net/sched: flower: Move filter handle initialization earlier
  net/sched: cls_api: Support hardware miss to tc action
  net/sched: Rename user cookie and act cookie
  sfc: fix builds without CONFIG_RTC_LIB
  sfc: clean up some inconsistent indentings
  net/mlx4_en: Introduce flexible array to silence overflow warning
  net: lan966x: Fix possible deadlock inside PTP
  net/ulp: Remove redundant ->clone() test in inet_clone_ulp().
  ...

1 files changed, 754 insertions, 0 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
new file mode 100644
index 000000000..9378fc79e
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gfx_v10.c
@@ -0,0 +1,754 @@
+/*
+ * Copyright 2019 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "amdgpu.h"
+#include "amdgpu_amdkfd.h"
+#include "gc/gc_10_1_0_offset.h"
+#include "gc/gc_10_1_0_sh_mask.h"
+#include "athub/athub_2_0_0_offset.h"
+#include "athub/athub_2_0_0_sh_mask.h"
+#include "oss/osssys_5_0_0_offset.h"
+#include "oss/osssys_5_0_0_sh_mask.h"
+#include "soc15_common.h"
+#include "v10_structs.h"
+#include "nv.h"
+#include "nvd.h"
+
+enum hqd_dequeue_request_type {
+	NO_ACTION = 0,
+	DRAIN_PIPE,
+	RESET_WAVES,
+	SAVE_WAVES
+};
+
+static void lock_srbm(struct amdgpu_device *adev, uint32_t mec, uint32_t pipe,
+			uint32_t queue, uint32_t vmid)
+{
+	mutex_lock(&adev->srbm_mutex);
+	nv_grbm_select(adev, mec, pipe, queue, vmid);
+}
+
+static void unlock_srbm(struct amdgpu_device *adev)
+{
+	nv_grbm_select(adev, 0, 0, 0, 0);
+	mutex_unlock(&adev->srbm_mutex);
+}
+
+static void acquire_queue(struct amdgpu_device *adev, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	uint32_t mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	uint32_t pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(adev, mec, pipe, queue_id, 0);
+}
+
+static uint64_t get_queue_mask(struct amdgpu_device *adev,
+			       uint32_t pipe_id, uint32_t queue_id)
+{
+	unsigned int bit = pipe_id * adev->gfx.mec.num_queue_per_pipe +
+			queue_id;
+
+	return 1ull << bit;
+}
+
+static void release_queue(struct amdgpu_device *adev)
+{
+	unlock_srbm(adev);
+}
+
+static void kgd_program_sh_mem_settings(struct amdgpu_device *adev, uint32_t vmid,
+					uint32_t sh_mem_config,
+					uint32_t sh_mem_ape1_base,
+					uint32_t sh_mem_ape1_limit,
+					uint32_t sh_mem_bases)
+{
+	lock_srbm(adev, 0, 0, 0, vmid);
+
+	WREG32_SOC15(GC, 0, mmSH_MEM_CONFIG, sh_mem_config);
+	WREG32_SOC15(GC, 0, mmSH_MEM_BASES, sh_mem_bases);
+	/* APE1 no longer exists on GFX9 */
+
+	unlock_srbm(adev);
+}
+
+static int kgd_set_pasid_vmid_mapping(struct amdgpu_device *adev, u32 pasid,
+					unsigned int vmid)
+{
+	/*
+	 * We have to assume that there is no outstanding mapping.
+	 * The ATC_VMID_PASID_MAPPING_UPDATE_STATUS bit could be 0 because
+	 * a mapping is in progress or because a mapping finished
+	 * and the SW cleared it.
+	 * So the protocol is to always wait & clear.
+	 */
+	uint32_t pasid_mapping = (pasid == 0) ? 0 : (uint32_t)pasid |
+			ATC_VMID0_PASID_MAPPING__VALID_MASK;
+
+	pr_debug("pasid 0x%x vmid %d, reg value %x\n", pasid, vmid, pasid_mapping);
+
+	pr_debug("ATHUB, reg %x\n", SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING) + vmid);
+	WREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING) + vmid,
+	       pasid_mapping);
+
+#if 0
+	/* TODO: uncomment this code when the hardware support is ready. */
+	while (!(RREG32(SOC15_REG_OFFSET(
+				ATHUB, 0,
+				mmATC_VMID_PASID_MAPPING_UPDATE_STATUS)) &
+		 (1U << vmid)))
+		cpu_relax();
+
+	pr_debug("ATHUB mapping update finished\n");
+	WREG32(SOC15_REG_OFFSET(ATHUB, 0,
+				mmATC_VMID_PASID_MAPPING_UPDATE_STATUS),
+	       1U << vmid);
+#endif
+
+	/* Mapping vmid to pasid also for IH block */
+	pr_debug("update mapping for IH block and mmhub");
+	WREG32(SOC15_REG_OFFSET(OSSSYS, 0, mmIH_VMID_0_LUT) + vmid,
+	       pasid_mapping);
+
+	return 0;
+}
+
+/* TODO - RING0 form of field is obsolete, seems to date back to SI
+ * but still works
+ */
+
+static int kgd_init_interrupts(struct amdgpu_device *adev, uint32_t pipe_id)
+{
+	uint32_t mec;
+	uint32_t pipe;
+
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	lock_srbm(adev, mec, pipe, 0, 0);
+
+	WREG32_SOC15(GC, 0, mmCPC_INT_CNTL,
+		CP_INT_CNTL_RING0__TIME_STAMP_INT_ENABLE_MASK |
+		CP_INT_CNTL_RING0__OPCODE_ERROR_INT_ENABLE_MASK);
+
+	unlock_srbm(adev);
+
+	return 0;
+}
+
+static uint32_t get_sdma_rlc_reg_offset(struct amdgpu_device *adev,
+				unsigned int engine_id,
+				unsigned int queue_id)
+{
+	uint32_t sdma_engine_reg_base[2] = {
+		SOC15_REG_OFFSET(SDMA0, 0,
+				 mmSDMA0_RLC0_RB_CNTL) - mmSDMA0_RLC0_RB_CNTL,
+		/* On gfx10, mmSDMA1_xxx registers are defined NOT based
+		 * on SDMA1 base address (dw 0x1860) but based on SDMA0
+		 * base address (dw 0x1260). Therefore use mmSDMA0_RLC0_RB_CNTL
+		 * instead of mmSDMA1_RLC0_RB_CNTL for the base address calc
+		 * below
+		 */
+		SOC15_REG_OFFSET(SDMA1, 0,
+				 mmSDMA1_RLC0_RB_CNTL) - mmSDMA0_RLC0_RB_CNTL
+	};
+
+	uint32_t retval = sdma_engine_reg_base[engine_id]
+		+ queue_id * (mmSDMA0_RLC1_RB_CNTL - mmSDMA0_RLC0_RB_CNTL);
+
+	pr_debug("RLC register offset for SDMA%d RLC%d: 0x%x\n", engine_id,
+			queue_id, retval);
+
+	return retval;
+}
+
+#if 0
+static uint32_t get_watch_base_addr(struct amdgpu_device *adev)
+{
+	uint32_t retval = SOC15_REG_OFFSET(GC, 0, mmTCP_WATCH0_ADDR_H) -
+			mmTCP_WATCH0_ADDR_H;
+
+	pr_debug("kfd: reg watch base address: 0x%x\n", retval);
+
+	return retval;
+}
+#endif
+
+static inline struct v10_compute_mqd *get_mqd(void *mqd)
+{
+	return (struct v10_compute_mqd *)mqd;
+}
+
+static inline struct v10_sdma_mqd *get_sdma_mqd(void *mqd)
+{
+	return (struct v10_sdma_mqd *)mqd;
+}
+
+static int kgd_hqd_load(struct amdgpu_device *adev, void *mqd,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t __user *wptr, uint32_t wptr_shift,
+			uint32_t wptr_mask, struct mm_struct *mm)
+{
+	struct v10_compute_mqd *m;
+	uint32_t *mqd_hqd;
+	uint32_t reg, hqd_base, data;
+
+	m = get_mqd(mqd);
+
+	pr_debug("Load hqd of pipe %d queue %d\n", pipe_id, queue_id);
+	acquire_queue(adev, pipe_id, queue_id);
+
+	/* HQD registers extend from CP_MQD_BASE_ADDR to CP_HQD_EOP_WPTR_MEM. */
+	mqd_hqd = &m->cp_mqd_base_addr_lo;
+	hqd_base = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
+
+	for (reg = hqd_base;
+	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
+		WREG32_SOC15_IP(GC, reg, mqd_hqd[reg - hqd_base]);
+
+
+	/* Activate doorbell logic before triggering WPTR poll. */
+	data = REG_SET_FIELD(m->cp_hqd_pq_doorbell_control,
+			     CP_HQD_PQ_DOORBELL_CONTROL, DOORBELL_EN, 1);
+	WREG32_SOC15(GC, 0, mmCP_HQD_PQ_DOORBELL_CONTROL, data);
+
+	if (wptr) {
+		/* Don't read wptr with get_user because the user
+		 * context may not be accessible (if this function
+		 * runs in a work queue). Instead trigger a one-shot
+		 * polling read from memory in the CP. This assumes
+		 * that wptr is GPU-accessible in the queue's VMID via
+		 * ATC or SVM. WPTR==RPTR before starting the poll so
+		 * the CP starts fetching new commands from the right
+		 * place.
+		 *
+		 * Guessing a 64-bit WPTR from a 32-bit RPTR is a bit
+		 * tricky. Assume that the queue didn't overflow. The
+		 * number of valid bits in the 32-bit RPTR depends on
+		 * the queue size. The remaining bits are taken from
+		 * the saved 64-bit WPTR. If the WPTR wrapped, add the
+		 * queue size.
+		 */
+		uint32_t queue_size =
+			2 << REG_GET_FIELD(m->cp_hqd_pq_control,
+					   CP_HQD_PQ_CONTROL, QUEUE_SIZE);
+		uint64_t guessed_wptr = m->cp_hqd_pq_rptr & (queue_size - 1);
+
+		if ((m->cp_hqd_pq_wptr_lo & (queue_size - 1)) < guessed_wptr)
+			guessed_wptr += queue_size;
+		guessed_wptr += m->cp_hqd_pq_wptr_lo & ~(queue_size - 1);
+		guessed_wptr += (uint64_t)m->cp_hqd_pq_wptr_hi << 32;
+
+		WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_LO,
+		       lower_32_bits(guessed_wptr));
+		WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_HI,
+		       upper_32_bits(guessed_wptr));
+		WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR,
+		       lower_32_bits((uint64_t)wptr));
+		WREG32_SOC15(GC, 0, mmCP_HQD_PQ_WPTR_POLL_ADDR_HI,
+		       upper_32_bits((uint64_t)wptr));
+		pr_debug("%s setting CP_PQ_WPTR_POLL_CNTL1 to %x\n", __func__,
+			 (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+		WREG32_SOC15(GC, 0, mmCP_PQ_WPTR_POLL_CNTL1,
+		       (uint32_t)get_queue_mask(adev, pipe_id, queue_id));
+	}
+
+	/* Start the EOP fetcher */
+	WREG32_SOC15(GC, 0, mmCP_HQD_EOP_RPTR,
+	       REG_SET_FIELD(m->cp_hqd_eop_rptr,
+			     CP_HQD_EOP_RPTR, INIT_FETCHER, 1));
+
+	data = REG_SET_FIELD(m->cp_hqd_active, CP_HQD_ACTIVE, ACTIVE, 1);
+	WREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE, data);
+
+	release_queue(adev);
+
+	return 0;
+}
+
+static int kgd_hiq_mqd_load(struct amdgpu_device *adev, void *mqd,
+			    uint32_t pipe_id, uint32_t queue_id,
+			    uint32_t doorbell_off)
+{
+	struct amdgpu_ring *kiq_ring = &adev->gfx.kiq.ring;
+	struct v10_compute_mqd *m;
+	uint32_t mec, pipe;
+	int r;
+
+	m = get_mqd(mqd);
+
+	acquire_queue(adev, pipe_id, queue_id);
+
+	mec = (pipe_id / adev->gfx.mec.num_pipe_per_mec) + 1;
+	pipe = (pipe_id % adev->gfx.mec.num_pipe_per_mec);
+
+	pr_debug("kfd: set HIQ, mec:%d, pipe:%d, queue:%d.\n",
+		 mec, pipe, queue_id);
+
+	spin_lock(&adev->gfx.kiq.ring_lock);
+	r = amdgpu_ring_alloc(kiq_ring, 7);
+	if (r) {
+		pr_err("Failed to alloc KIQ (%d).\n", r);
+		goto out_unlock;
+	}
+
+	amdgpu_ring_write(kiq_ring, PACKET3(PACKET3_MAP_QUEUES, 5));
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_MAP_QUEUES_QUEUE_SEL(0) | /* Queue_Sel */
+			  PACKET3_MAP_QUEUES_VMID(m->cp_hqd_vmid) | /* VMID */
+			  PACKET3_MAP_QUEUES_QUEUE(queue_id) |
+			  PACKET3_MAP_QUEUES_PIPE(pipe) |
+			  PACKET3_MAP_QUEUES_ME((mec - 1)) |
+			  PACKET3_MAP_QUEUES_QUEUE_TYPE(0) | /*queue_type: normal compute queue */
+			  PACKET3_MAP_QUEUES_ALLOC_FORMAT(0) | /* alloc format: all_on_one_pipe */
+			  PACKET3_MAP_QUEUES_ENGINE_SEL(1) | /* engine_sel: hiq */
+			  PACKET3_MAP_QUEUES_NUM_QUEUES(1)); /* num_queues: must be 1 */
+	amdgpu_ring_write(kiq_ring,
+			  PACKET3_MAP_QUEUES_DOORBELL_OFFSET(doorbell_off));
+	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_lo);
+	amdgpu_ring_write(kiq_ring, m->cp_mqd_base_addr_hi);
+	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_lo);
+	amdgpu_ring_write(kiq_ring, m->cp_hqd_pq_wptr_poll_addr_hi);
+	amdgpu_ring_commit(kiq_ring);
+
+out_unlock:
+	spin_unlock(&adev->gfx.kiq.ring_lock);
+	release_queue(adev);
+
+	return r;
+}
+
+static int kgd_hqd_dump(struct amdgpu_device *adev,
+			uint32_t pipe_id, uint32_t queue_id,
+			uint32_t (**dump)[2], uint32_t *n_regs)
+{
+	uint32_t i = 0, reg;
+#define HQD_N_REGS 56
+#define DUMP_REG(addr) do {				\
+		if (WARN_ON_ONCE(i >= HQD_N_REGS))	\
+			break;				\
+		(*dump)[i][0] = (addr) << 2;		\
+		(*dump)[i++][1] = RREG32_SOC15_IP(GC, addr);		\
+	} while (0)
+
+	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	acquire_queue(adev, pipe_id, queue_id);
+
+	for (reg = SOC15_REG_OFFSET(GC, 0, mmCP_MQD_BASE_ADDR);
+	     reg <= SOC15_REG_OFFSET(GC, 0, mmCP_HQD_PQ_WPTR_HI); reg++)
+		DUMP_REG(reg);
+
+	release_queue(adev);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static int kgd_hqd_sdma_load(struct amdgpu_device *adev, void *mqd,
+			     uint32_t __user *wptr, struct mm_struct *mm)
+{
+	struct v10_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	unsigned long end_jiffies;
+	uint32_t data;
+	uint64_t data64;
+	uint64_t __user *wptr64 = (uint64_t __user *)wptr;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL,
+		m->sdmax_rlcx_rb_cntl & (~SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK));
+
+	end_jiffies = msecs_to_jiffies(2000) + jiffies;
+	while (true) {
+		data = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_CONTEXT_STATUS);
+		if (data & SDMA0_RLC0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL_OFFSET,
+	       m->sdmax_rlcx_doorbell_offset);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_doorbell, SDMA0_RLC0_DOORBELL,
+			     ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL, data);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR,
+				m->sdmax_rlcx_rb_rptr);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_HI,
+				m->sdmax_rlcx_rb_rptr_hi);
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_MINOR_PTR_UPDATE, 1);
+	if (read_user_wptr(mm, wptr64, data64)) {
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR,
+		       lower_32_bits(data64));
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR_HI,
+		       upper_32_bits(data64));
+	} else {
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR,
+		       m->sdmax_rlcx_rb_rptr);
+		WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_WPTR_HI,
+		       m->sdmax_rlcx_rb_rptr_hi);
+	}
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_MINOR_PTR_UPDATE, 0);
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_BASE, m->sdmax_rlcx_rb_base);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_BASE_HI,
+			m->sdmax_rlcx_rb_base_hi);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_ADDR_LO,
+			m->sdmax_rlcx_rb_rptr_addr_lo);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_ADDR_HI,
+			m->sdmax_rlcx_rb_rptr_addr_hi);
+
+	data = REG_SET_FIELD(m->sdmax_rlcx_rb_cntl, SDMA0_RLC0_RB_CNTL,
+			     RB_ENABLE, 1);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL, data);
+
+	return 0;
+}
+
+static int kgd_hqd_sdma_dump(struct amdgpu_device *adev,
+			     uint32_t engine_id, uint32_t queue_id,
+			     uint32_t (**dump)[2], uint32_t *n_regs)
+{
+	uint32_t sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev,
+			engine_id, queue_id);
+	uint32_t i = 0, reg;
+#undef HQD_N_REGS
+#define HQD_N_REGS (19+6+7+10)
+
+	*dump = kmalloc(HQD_N_REGS*2*sizeof(uint32_t), GFP_KERNEL);
+	if (*dump == NULL)
+		return -ENOMEM;
+
+	for (reg = mmSDMA0_RLC0_RB_CNTL; reg <= mmSDMA0_RLC0_DOORBELL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = mmSDMA0_RLC0_STATUS; reg <= mmSDMA0_RLC0_CSA_ADDR_HI; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = mmSDMA0_RLC0_IB_SUB_REMAIN;
+	     reg <= mmSDMA0_RLC0_MINOR_PTR_UPDATE; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+	for (reg = mmSDMA0_RLC0_MIDCMD_DATA0;
+	     reg <= mmSDMA0_RLC0_MIDCMD_CNTL; reg++)
+		DUMP_REG(sdma_rlc_reg_offset + reg);
+
+	WARN_ON_ONCE(i != HQD_N_REGS);
+	*n_regs = i;
+
+	return 0;
+}
+
+static bool kgd_hqd_is_occupied(struct amdgpu_device *adev,
+				uint64_t queue_address, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	uint32_t act;
+	bool retval = false;
+	uint32_t low, high;
+
+	acquire_queue(adev, pipe_id, queue_id);
+	act = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE);
+	if (act) {
+		low = lower_32_bits(queue_address >> 8);
+		high = upper_32_bits(queue_address >> 8);
+
+		if (low == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE) &&
+		   high == RREG32_SOC15(GC, 0, mmCP_HQD_PQ_BASE_HI))
+			retval = true;
+	}
+	release_queue(adev);
+	return retval;
+}
+
+static bool kgd_hqd_sdma_is_occupied(struct amdgpu_device *adev, void *mqd)
+{
+	struct v10_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t sdma_rlc_rb_cntl;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	sdma_rlc_rb_cntl = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL);
+
+	if (sdma_rlc_rb_cntl & SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK)
+		return true;
+
+	return false;
+}
+
+static int kgd_hqd_destroy(struct amdgpu_device *adev, void *mqd,
+				enum kfd_preempt_type reset_type,
+				unsigned int utimeout, uint32_t pipe_id,
+				uint32_t queue_id)
+{
+	enum hqd_dequeue_request_type type;
+	unsigned long end_jiffies;
+	uint32_t temp;
+	struct v10_compute_mqd *m = get_mqd(mqd);
+
+	if (amdgpu_in_reset(adev))
+		return -EIO;
+
+#if 0
+	unsigned long flags;
+	int retry;
+#endif
+
+	acquire_queue(adev, pipe_id, queue_id);
+
+	if (m->cp_hqd_vmid == 0)
+		WREG32_FIELD15(GC, 0, RLC_CP_SCHEDULERS, scheduler1, 0);
+
+	switch (reset_type) {
+	case KFD_PREEMPT_TYPE_WAVEFRONT_DRAIN:
+		type = DRAIN_PIPE;
+		break;
+	case KFD_PREEMPT_TYPE_WAVEFRONT_RESET:
+		type = RESET_WAVES;
+		break;
+	case KFD_PREEMPT_TYPE_WAVEFRONT_SAVE:
+		type = SAVE_WAVES;
+		break;
+	default:
+		type = DRAIN_PIPE;
+		break;
+	}
+
+#if 0 /* Is this still needed? */
+	/* Workaround: If IQ timer is active and the wait time is close to or
+	 * equal to 0, dequeueing is not safe. Wait until either the wait time
+	 * is larger or timer is cleared. Also, ensure that IQ_REQ_PEND is
+	 * cleared before continuing. Also, ensure wait times are set to at
+	 * least 0x3.
+	 */
+	local_irq_save(flags);
+	preempt_disable();
+	retry = 5000; /* wait for 500 usecs at maximum */
+	while (true) {
+		temp = RREG32(mmCP_HQD_IQ_TIMER);
+		if (REG_GET_FIELD(temp, CP_HQD_IQ_TIMER, PROCESSING_IQ)) {
+			pr_debug("HW is processing IQ\n");
+			goto loop;
+		}
+		if (REG_GET_FIELD(temp, CP_HQD_IQ_TIMER, ACTIVE)) {
+			if (REG_GET_FIELD(temp, CP_HQD_IQ_TIMER, RETRY_TYPE)
+					== 3) /* SEM-rearm is safe */
+				break;
+			/* Wait time 3 is safe for CP, but our MMIO read/write
+			 * time is close to 1 microsecond, so check for 10 to
+			 * leave more buffer room
+			 */
+			if (REG_GET_FIELD(temp, CP_HQD_IQ_TIMER, WAIT_TIME)
+					>= 10)
+				break;
+			pr_debug("IQ timer is active\n");
+		} else
+			break;
+loop:
+		if (!retry) {
+			pr_err("CP HQD IQ timer status time out\n");
+			break;
+		}
+		ndelay(100);
+		--retry;
+	}
+	retry = 1000;
+	while (true) {
+		temp = RREG32(mmCP_HQD_DEQUEUE_REQUEST);
+		if (!(temp & CP_HQD_DEQUEUE_REQUEST__IQ_REQ_PEND_MASK))
+			break;
+		pr_debug("Dequeue request is pending\n");
+
+		if (!retry) {
+			pr_err("CP HQD dequeue request time out\n");
+			break;
+		}
+		ndelay(100);
+		--retry;
+	}
+	local_irq_restore(flags);
+	preempt_enable();
+#endif
+
+	WREG32_SOC15(GC, 0, mmCP_HQD_DEQUEUE_REQUEST, type);
+
+	end_jiffies = (utimeout * HZ / 1000) + jiffies;
+	while (true) {
+		temp = RREG32_SOC15(GC, 0, mmCP_HQD_ACTIVE);
+		if (!(temp & CP_HQD_ACTIVE__ACTIVE_MASK))
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("cp queue preemption time out.\n");
+			release_queue(adev);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	release_queue(adev);
+	return 0;
+}
+
+static int kgd_hqd_sdma_destroy(struct amdgpu_device *adev, void *mqd,
+				unsigned int utimeout)
+{
+	struct v10_sdma_mqd *m;
+	uint32_t sdma_rlc_reg_offset;
+	uint32_t temp;
+	unsigned long end_jiffies = (utimeout * HZ / 1000) + jiffies;
+
+	m = get_sdma_mqd(mqd);
+	sdma_rlc_reg_offset = get_sdma_rlc_reg_offset(adev, m->sdma_engine_id,
+					    m->sdma_queue_id);
+
+	temp = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL);
+	temp = temp & ~SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK;
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL, temp);
+
+	while (true) {
+		temp = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_CONTEXT_STATUS);
+		if (temp & SDMA0_RLC0_CONTEXT_STATUS__IDLE_MASK)
+			break;
+		if (time_after(jiffies, end_jiffies)) {
+			pr_err("SDMA RLC not idle in %s\n", __func__);
+			return -ETIME;
+		}
+		usleep_range(500, 1000);
+	}
+
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_DOORBELL, 0);
+	WREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL,
+		RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_CNTL) |
+		SDMA0_RLC0_RB_CNTL__RB_ENABLE_MASK);
+
+	m->sdmax_rlcx_rb_rptr = RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR);
+	m->sdmax_rlcx_rb_rptr_hi =
+		RREG32(sdma_rlc_reg_offset + mmSDMA0_RLC0_RB_RPTR_HI);
+
+	return 0;
+}
+
+static bool get_atc_vmid_pasid_mapping_info(struct amdgpu_device *adev,
+					uint8_t vmid, uint16_t *p_pasid)
+{
+	uint32_t value;
+
+	value = RREG32(SOC15_REG_OFFSET(ATHUB, 0, mmATC_VMID0_PASID_MAPPING)
+		     + vmid);
+	*p_pasid = value & ATC_VMID0_PASID_MAPPING__PASID_MASK;
+
+	return !!(value & ATC_VMID0_PASID_MAPPING__VALID_MASK);
+}
+
+static int kgd_wave_control_execute(struct amdgpu_device *adev,
+					uint32_t gfx_index_val,
+					uint32_t sq_cmd)
+{
+	uint32_t data = 0;
+
+	mutex_lock(&adev->grbm_idx_mutex);
+
+	WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX, gfx_index_val);
+	WREG32_SOC15(GC, 0, mmSQ_CMD, sq_cmd);
+
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		INSTANCE_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SA_BROADCAST_WRITES, 1);
+	data = REG_SET_FIELD(data, GRBM_GFX_INDEX,
+		SE_BROADCAST_WRITES, 1);
+
+	WREG32_SOC15(GC, 0, mmGRBM_GFX_INDEX, data);
+	mutex_unlock(&adev->grbm_idx_mutex);
+
+	return 0;
+}
+
+static void set_vm_context_page_table_base(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t page_table_base)
+{
+	if (!amdgpu_amdkfd_is_kfd_vmid(adev, vmid)) {
+		pr_err("trying to set page table base for wrong VMID %u\n",
+		       vmid);
+		return;
+	}
+
+	/* SDMA is on gfxhub as well for Navi1* series */
+	adev->gfxhub.funcs->setup_vm_pt_regs(adev, vmid, page_table_base);
+}
+
+static void program_trap_handler_settings(struct amdgpu_device *adev,
+		uint32_t vmid, uint64_t tba_addr, uint64_t tma_addr)
+{
+	lock_srbm(adev, 0, 0, 0, vmid);
+
+	/*
+	 * Program TBA registers
+	 */
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_LO),
+			lower_32_bits(tba_addr >> 8));
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TBA_HI),
+			upper_32_bits(tba_addr >> 8) |
+			(1 << SQ_SHADER_TBA_HI__TRAP_EN__SHIFT));
+
+	/*
+	 * Program TMA registers
+	 */
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_LO),
+			lower_32_bits(tma_addr >> 8));
+	WREG32(SOC15_REG_OFFSET(GC, 0, mmSQ_SHADER_TMA_HI),
+			upper_32_bits(tma_addr >> 8));
+
+	unlock_srbm(adev);
+}
+
+const struct kfd2kgd_calls gfx_v10_kfd2kgd = {
+	.program_sh_mem_settings = kgd_program_sh_mem_settings,
+	.set_pasid_vmid_mapping = kgd_set_pasid_vmid_mapping,
+	.init_interrupts = kgd_init_interrupts,
+	.hqd_load = kgd_hqd_load,
+	.hiq_mqd_load = kgd_hiq_mqd_load,
+	.hqd_sdma_load = kgd_hqd_sdma_load,
+	.hqd_dump = kgd_hqd_dump,
+	.hqd_sdma_dump = kgd_hqd_sdma_dump,
+	.hqd_is_occupied = kgd_hqd_is_occupied,
+	.hqd_sdma_is_occupied = kgd_hqd_sdma_is_occupied,
+	.hqd_destroy = kgd_hqd_destroy,
+	.hqd_sdma_destroy = kgd_hqd_sdma_destroy,
+	.wave_control_execute = kgd_wave_control_execute,
+	.get_atc_vmid_pasid_mapping_info =
+			get_atc_vmid_pasid_mapping_info,
+	.set_vm_context_page_table_base = set_vm_context_page_table_base,
+	.program_trap_handler_settings = program_trap_handler_settings,
+};